Forecasting Iran’s Energy Demand Using Cuckoo Optimization Algorithm

Table 4

Comparison of the COA estimation models for energy consumption in the testing period (2009–2013).

Years	2009	2010	2011	2012	2013	Average
Actual data^a	1158.311	1134.874	1184.621	1182.055	1229.75	—
COA-linear model 1	1146.673	1167.896	1175.386	1199.447	1221.124	—
Relative error (%)	−1.01495	2.8275	−0.78565	1.4499	−0.7063	1.3568
COA-linear model 2	1152.347	1140.204	1175.386	1185.487	1192.07
Relative error (%)	−0.5175	0.4674	−0.7856	0.2894	−3.1608	1.0441
COA-linear model 3	1150.078	1161.654	1166.08	1198.539	1217.719
Relative error (%)	−0.7159	2.3053	−1.5900	1.3753	−0.9879	1.3948
COA-linear model 4	1160.292	1134.983	1179.132	1196.383	1208.753
Relative error (%)	0.1707	0.0096	−0.4655	1.1975	−1.7370	0.716
COA-power model 1	1137.48	1158.93	1165.626	1204.1	1235.424
Relative error (%)	−1.8313	2.0757	−1.6295	1.8308	0.4593	1.5653
COA-power model 2	1152.574	1125.201	1124.201	1174.252	1195.248
Relative error (%)	−0.49772	−0.9493	−5.3744	−0.6645	−2.8865	2.0745
COA-power model 3	1161.99	1160.065	1163.924	1217.379	1220.67
Relative error (%)	0.3169	2.1715	−1.7782	2.9015	−0.7438	1.5824
COA-power model 4	1129.422	1208.98	1265.613	1230.657	1216.13
Relative error (%)	−2.5578	6.1296	6.3994	3.9492	−1.1198	4.0311

^aActual data are in million barrel oil equivalent (Mboe).